Asian Spine Journal

Clinical
850 Ayush Sharma et Study
al. Asian Spine J 2016;10(5):850-856
Asian Spine J 2016;10(5):850-856 • https://doi.org/10.4184/asj.2016.10.5.850

Magnetic Resonance Imaging and GeneXpert:

A Rapid and Accurate Diagnostic Tool for the
Management of Tuberculosis of the Spine
Ayush Sharma1,2, Harvinder Singh Chhabra2, Rajat Mahajan2, Tarun Chabra2, Sahil Batra2
1
Department of Orthopedic and Spine Surgery, Dr. Babasaheb Ambedkar Central Railway Hospital, Mumbai, India
2
Department of Spine Services Indian Spinal Injuries Center, New Delhi, India

Study Design: Retrospective study.

Purpose: The aim of this study was to analyze various diagnostic tools, including GeneXpert, for the management of tuberculosis of
the spine.
Overview of Literature: Traditional diagnostic methods of microscopy, histology, and culture have low sensitivity and specificity for
the management of tuberculosis of the spine.
Methods: Of the 262 treated cases of spinal tuberculosis, data on 1 year follow-up was available for 217 cases. Of these, only 145
cases with a confirmed diagnosis were selected for retrospective analysis.
Results: In 145 of the 217 patients (66.80%), diagnosis was confirmed on the basis of a culture. Of the 145 patients with a confirmed
diagnosis, 98 (66.20%) patients were diagnosed on the basis of clinical presentation, whereas 123 (84.8%) exhibited a typical mag-
netic resonance imaging (MRI) picture. In 99 surgically treated patients, the diagnosis was confirmed on the basis of an intraoperative
tissue biopsy. Among the 46 patients treated conservatively, 35 underwent a transpedicular biopsy, 4 patients underwent computed
tomography-guided biopsy, 6 patients were diagnosed on the basis of material obtained from a cold abscess, and 1 patient underwent
an open biopsy. The sensitivity of the culture for the detection of Mycobacterium tuberculosis was 66.80% (145/217) in our patients.
Among the cases in which GeneXpert was used, the sensitivity for the detection of Mycobacterium tuberculosis was 93.4% (43/46).
Moreover, the sensitivity of GeneXpert to detect rifampicin resistance was 100% (7/7) in our study.
Conclusions: Majority of the patients with tuberculosis of the spine can be diagnosed on the basis of a typical radiological presen-
tation via MRI. In our study, 84.8% cases exhibited typical MRI findings. For patients presenting with atypical MRI features, a rapid
and accurate diagnosis is possible by combining GeneXpert with MRI. The combined use of MRI and GeneXpert is a rapid and highly
sensitive tool to diagnose tuberculosis and rifampicin resistance in patients with tuberculosis of the spine. Furthermore, we achieved
a 97.9% sensitivity for the detection of Mycobacterium tuberculosis and 100% sensitivity for the detection of rifampicin resistance in
our study.

Keywords: GeneXpert; Tuberculosis; Diagnosis; Spine

Received Jan 10, 2016; Revised Feb 9, 2016; Accepted Feb 24, 2016
Corresponding author: Ayush Sharma
Doctor’s quarter no 6, Dr B R Ambedker Central Railway Hospital, Byculla east, Mumbai, India 400027
Tel: +91-90-0454-9623, E-mail: drayush@gmail.com

ASJ
Copyright Ⓒ 2016 by Korean Society of Spine Surgery
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/)
which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Asian Spine Journal • pISSN 1976-1902 eISSN 1976-7846 • www.asianspinejournal.org
 Asian Spine Journal
Introduction
Musculoskeletal affection is observed in 4% of all cases
with tuberculosis; 50% of which involve the spine [1],
which is the most common form of skeletal tuberculosis.
Currently, the diagnosis of tuberculosis of the spine is
primarily based on clinico-radiological observations. A
typical presentation of tuberculosis of the spine consists of
pain during movement with a localized deformity in the
back that is tender following percussion as well as other
typical systemic symptoms of active tuberculosis (i.e.,
night cries, malaise, weight loss, loss of appetite, night
sweats, and a rise in temperature in the evening). More-
over, patients may or may not have a neurological deficit,
which can be the first symptom in rare cases [1]. Micro-
scopic confirmation using Ziehl–Nielsen staining remains
a popular diagnostic method because of its simplicity and
cost-effectiveness; however, it has a low sensitivity and
requires 10,000–100,000 bacilli/mL in clinical specimens
to be positive [2,3]. Mycobacterium tuberculosis (Mtb)
culture is the gold standard method for the diagnosis of
tuberculosis, but it also has various limitations, includ-
ing a required 6–8 week period of growth because of the
slow replication rate of the bacteria; these results are often
negative as it requires 10–100 bacilli/mL (live bacilli) in
clinical specimens to achieve culture positive results [4,5].
Magnetic resonance imaging (MRI) is a better diagnostic
method than radiography [6]. Marrow edema, endplate
disruption, paravertebral soft tissue formation, subliga-
mentous collections, and a high signal of the intervertebral
disc on T2-weighted are typical MRI features with good
to excellent sensitivity for spinal tuberculosis. Overall, the
sensitivity and specificity of MRI for spinal tuberculosis
are 100% and 88.2%, respectively [7,8]. MRI findings with
a high sensitivity and specificity include the disruption of
the end-plate (100% and 81.4%, respectively), paraverte-
bral soft-tissue shadow (96.8% and 85.3%, respectively),
and high signal intensity of the intervertebral disc on
the T2-weighted image (80.6% and 82.4%, respectively).
Atypical presentation primarily includes discrete foci of
spinal involvement with intervening normal vertebrae
and no evidence of a connecting soft tissue abscess or any
other MRI features typical of tuberculosis as discussed
above (i.e., the involvement of only the posterior column
of the spine without end plate involvement and multiple
skip lesions without a soft tissue shadow). In patients with
an atypical clinical and MRI presentation, further inves-
GeneXpert: a rapid and accurate diagnostic tool 851
tigation in the form of an open or precautionary biopsy
can be performed to confirm the diagnosis. Although the
specimen adequacy is higher for an open biopsy than for a
percutaneous biopsy, the similarity ratio between the ini-
tial radiological and final pathological diagnosis of both
techniques are favorable (71.4% for the open biopsy and
69.2% for the percutaneous biopsy) [9]. Recent techniques,
such as polymerase chain reaction (PCR) and GeneXpert
provide improved accuracy over microscopy and are more
rapid than bacterial cultures. The GeneXpert test has a sen-
sitivity of 95.6% and a specificity of 96.2% for diagnosis of
spinal tuberculosis [10].
Additionally, GeneXpert is more likely to detect Mtb
DNA than traditional PCR, with the added advantage
of also determining rifampicin resistance [11]. A delay
in diagnosis and the failure to detect drug resistance are
major hurdles involved in the treatment of tuberculosis of
the spine even today. The aim of this study was to analyze
various diagnostic tools for the management of tubercu-
losis of the spine, particularly the use of rapid diagnostic
tools, such as GeneXpert and MRI.
Materials and Methods
All cases with tuberculosis of the spine treated at the In-
dian Spinal Injuries Center (New Delhi), a tertiary care
spine and rehabilitation center, between October 2012
and December 2014 were retrospectively analyzed. Oc-
tober 2012 was used as the cutoff as this was the time our
center initiated the use of GeneXpert as a diagnostic tool
for tuberculosis of the spine. Of the 262 cases during this
period, data on a follow-up of more than 1 year was avail-
able for 217 cases. Of these 217 cases, only 145 (66.80%)
cases were bacteriologically confirmed via a bacterial
culture and were included in the present study. One of the
authors independently analyzed the MRI films, reports,
and history of all the patients and categorized them into
typical and atypical cases (Figs. 1A, 2A). All cases that did
not exhibit a typical presentation as discussed above were
categorized as “atypical.” Diagnostic biopsy methods and
the use of GeneXpert were also studied. Data were ana-
lyzed according to the documented American Spinal In-
jury Association (AIS) impairment scale [12] at the time
of presentation and at the final follow-up to determine the
extent of neurological recovery. All cases that presented
with a neurological deficit were divided into three groups:
(1) completely improved (AIS grade at the final follow-up
852 Ayush Sharma et al. Asian Spine J 2016;10(5):850-856

A C
Fig. 1. An atypical case of tuberculosis of the spine. A 22-year-old female patient presented with only back pain. (A)
Atypical magnetic resonance imaging (MRI) presentation via the T2-weighted image. (B) Percutaneous intradiscal bi-
opsy was performed from the lumber 3–4 disc, and GeneXpert was used for diagnostic confirmation. (C) Follow-up MRI
revealed the T2-weighted image after conservative treatment showing resolution of the disease.

B
Fig. 2. A typical case of tuberculosis of the spine. A
43-year-old male patient presented with back pain, ten-
derness in the thoracic spine, and neurological deficit. (A)
Typical magnetic resonance imaging picture of tuberculo-
sis of the spine on the T2-weighted image. (B) Diagnosis
was established on the basis of an intraoperative tissue
sample in this case. The figure shows the postoperative
X-ray following surgery.

A
 Asian Spine Journal
was E), (2) partially improved (AIS grade changed to B, C,
or D but not E), and (3) no improvement (AIS grading at
presentation and final follow-up remained unchanged).
Results
The average patient age was 45.8±19.1 years, and 71
male and 74 female patients were included. A total of 98
(66.20%) patients exhibited the typical clinical presenta-
tion, whereas 123 (84.8%) presented with the typical MRI
findings as discussed above. There were 99 patients who
were surgically treated, whereas 46 patients were man-
aged conservatively (Figs. 1C, 2B). In all of the surgically
treated cases, the diagnosis was made on the basis of an
intraoperative tissue biopsy obtained from the diseased
vertebra and intervening disc space. In the 46 patients
managed conservatively, 35 patients underwent a trans-
pedicular or intradiscal percutaneous biopsy (Fig. 1B), 4
patients underwent a computed tomography (CT)-guided
biopsy; 6 patients were diagnosed on the basis of material
obtained from cold abscess and 1 patient underwent open
biopsy (Fig. 3).
GeneXpert was used in 46 cases, from which 23 tissue
samples were obtained from an intraoperative biopsy, 18
were attained via a transpedicular biopsy, 3 were obtained
from the aspiration of a cold abscess, and 1 was obtained
from CT-guided and open biopsies. Of the 43 cases that
tested positive using GeneXpert, seven also tested positive
Fig. 3. A graph showing the type of biopsy (X-axis) and number of patients (Y-axis). CT, computed tomography.
GeneXpert: a rapid and accurate diagnostic tool 853
for rifampicin resistance (Fig. 4). The patients that tested
positive for rifampicin resistance were further confirmed
by a drug sensitivity test. Although the sensitivity of Gen-
eXpert for the detection of rifampicin resistance was 100%
(7/7) in our study, the overall sensitivity for the detection
of Mtb was 93.4% (43/46). In addition, GeneXpert was
used in all 22 patients with an atypical MRI to confirm
the diagnosis. In the 44 patients who were neurologically
intact at presentation, no neurological deterioration was
observed at the final follow-up after the administration of
appropriate conservative or surgical treatment. Following
treatment, among the 101 patients who presented with a
neurological deficit according to the AIS grading system,
70 (69.30%) exhibited a complete neurological improve-
ment, 21 (20.7%) underwent a partial improvement, and
10 (9.9%) patients showed no improvement at the final
follow-up (Fig. 5).
Discussion
A delay in both the diagnosis and initiation of treatment
as well as the failure to recognize cases of drug resistance
could have an adverse effect on the prognosis of patients
with tuberculosis of the spine [13,14]. Traditional diag-
nostic methods are based on the typical clinical features
followed by a bacteriological confirmation via positive
histology and culture. Moreover, the traditional methods
of microscopy, histology, and culture have a low sensitivity
 854 Ayush Sharma et al. Asian Spine J 2016;10(5):850-856

Fig. 4. A graph showing whether GeneXpert was used, if Mycobacterium tuberculosis was detected, the presence of
rifampicin resistance (X-axis), and the number of patients (Y-axis).

Fig. 5. A graph of the neurological improvement (X-axis) and the number of patients (Y-axis).

and specificity. The sensitivity of histology to confirm a
diagnosis of spinal tuberculosis has been reported to be
approximately 60%. In addition, the incidence of positive
cultures for acid-fast bacilli in osteoarticular tuberculous
lesions has been reported to be between 40% and 88%
[15-18]. In our present study, to confirm the diagnosis of
tuberculosis of the spine, the sensitivity of the culture was
66.80% (145/217) in the treated cases. Although there is
no data regarding the incidence of atypical clinical fea-
tures in patients with tuberculosis of the spine, 47 (32.4%)
patients in our study presented with atypical clinical
features, suggesting that one in every three patient with
tuberculosis can exhibit an atypical clinical presentation.
Currently, MRI is the most popular tool used for the di-
agnosis of tuberculosis of the spine with good to excellent
sensitivity [7,8]. However, atypical MRI features have been
 Asian Spine Journal
reported in literature to range from 10% to 25% in cases
of tuberculosis of the spine [19]. Polley and Dunn [19]
reported atypical MRI presentation in 16.3% of the 98 pa-
tients from a single surgeon series, with a higher incidence
of neurological symptoms in these cases. In our study, an
atypical MRI picture was observed in 22 patients (15.17%).
In addition, GeneXpert was used in 46 cases, including 23
tissue samples obtained from an intraoperative biopsy, 18
acquired via a transpedicular biopsy, 3 from the aspiration
of a cold abscess, and 1 from a CT-guided and open biopsy.
Although the sensitivity of GeneXpert was very high at
95.8% (23/24) for tissue samples taken during surgery or
open biopsy from diseased vertebra and disc material,
the sensitivity was reduced to 90.90% (20/22) for tissue
samples obtained via a percutaneous biopsy. Moreover,
the overall sensitivity was 93.4% (43/46), and the sensi-
tivity of GeneXpert to detect rifampicin resistance was
100% (7/7). Furthermore, the rate of rifampicin resistance
was 4.8% (7/145) in our study. In all cases exhibiting an
atypical MRI presentation, GeneXpert was used to make
a rapid and accurate diagnosis. By combining MRI with
GeneXpert, we were able to achieve a considerably high
sensitivity of 97.9% for the detection of Mtb in cases of
tuberculosis of the spine. Recent literature on GeneXpert
supports our findings that the sensitivity for the detection
of Mtb and rifampicin resistance using this method can be
as high as 95% and 100%, respectively [10,11]. However,
there are no references in the literature regarding the ac-
curacy of combining MRI and GeneXpert for the diagno-
sis of tuberculosis of the spine. Although none of the cases
that were neurologically intact at presentation exhibited
any neurological deterioration following treatment, 70
patients demonstrated complete improvement, 21 patients
partially improved, and 10 patients showed no improve-
ment at the final follow-up. Therefore, no improvement
overall was only observed in 6.8% (10/145) of the patients,
with 14.4% (21/145) exhibiting a partial improvement,
and 79.62% (114/145) with a complete neurological
improvement following treatment using this diagnostic
protocol.
Conclusions
An accurate clinical diagnosis of tuberculosis of the spine
requires an extremely high degree of clinical suspicion as
one in every three patient can exhibit an atypical clinical
presentation. In our present study, 32.4% of patients had
GeneXpert: a rapid and accurate diagnostic tool 855
an atypical clinical presentation. A culture of acid-fast
bacilli in osteoarticular tuberculous lesions remains the
gold standard diagnostic test, but it is far from being an
ideal screening tool to diagnose tuberculosis of the spine
because of its low sensitivity (66.80% in our study). More-
over, MRI is a good screening tool, but 15.17% of cases in
our study exhibited an atypical MRI presentation. In con-
trast, the sensitivity of GeneXpert for detecting Mtb and
rifampicin resistance is excellent. In our study, GeneXpert
had a sensitivity of 93.4% for detecting Mtb and was had
a sensitivity of 100% for predicting rifampicin resistance.
Combining MRI with GeneXpert, particularly in cases
with an atypical presentation provides a rapid and highly
sensitive diagnosis tool to detect both Mtb and rifampicin
resistance in patients with tuberculosis of the spine. The
use of GeneXpert in conjunction with MRI was 97.9%
sensitive for the diagnosis of tuberculosis of the spine in
our study.
Conflict of Interest
No potential conflict of interest relevant to this article was
reported.
Reference
1. Tuli SM. Tuberculosis of the skeletal system: bones,
joints, spine and bursal sheaths. New Delhi: Jaypee
Brothers Pvt. Ltd.; 2010.
2. Kumar M, Kumar R, Srivastava AK, et al. Sensitivity
of PCR IS6110 in relation to culture and staining in
Pott’s disease. Indian J Neurosurg 2013;2:46-51.
3. Singh UB, Bhanu NV, Suresh VN, Arora J, Rana T,
Seth P. Utility of polymerase chain reaction in diag-
nosis of tuberculosis from samples of bone marrow
aspirate. Am J Trop Med Hyg 2006;75:960-3.
4. Negi SS, Gupta S, Khare S, Lal S. Comparison of vari-
ous microbiological tests including polymerase chain
reaction for the diagnosis of osteoarticular tubercu-
losis. Indian J Med Microbiol 2005;23:245-8.
5. Pandey V, Chawla K, Acharya K, Rao S, Rao S. The
role of polymerase chain reaction in the management
of osteoarticular tuberculosis. Int Orthop 2009;33:
801-5.
6. Griffith JF, Kumta SM, Leung PC, Cheng JC, Chow
LT, Metreweli C. Imaging of musculoskeletal tuber-
culosis: a new look at an old disease. Clin Orthop
 856 Ayush Sharma et al.
Relat Res 2002;(398):32-9..
7. Danchaivijitr N, Temram S, Thepmongkhol K,
Chiewvit P. Diagnostic accuracy of MR imaging in
tuberculous spondylitis. J Med Assoc Thai 2007;90:
1581-9.
8. Jain AK, Sreenivasan R, Saini NS, Kumar S, Jain S,
Dhammi IK. Magnetic resonance evaluation of tu-
bercular lesion in spine. Int Orthop 2012;36:261-9.
9. Yapici F, Atici Y, Balioglu MB, et al. A comparison of
two techniques: open and percutaneous biopsies of
thoracolumbar vertebral body lesions. J Craniover-
tebr Junction Spine 2015;6:36-9.
10. Held M, Laubscher M, Zar HJ, Dunn RN. GeneXpert
polymerase chain reaction for spinal tuberculosis: an
accurate and rapid diagnostic test. Bone Joint J 2014;
96B:1366-9.
11. Yagmur G, Albayrak N, Das T, Yildirim M, Ozgun A,
Buyuk Y. Comparison of two different real-time PCR
systems in postmortem diagnosis of tuberculosis in
paraffin-embedded tissues. Mikrobiyol Bul 2014;48:
577-84.
12. Maynard FM Jr, Bracken MB, Creasey G, et al. Inter-
Asian Spine J 2016;10(5):850-856
national Standards for Neurological and Functional
Classification of Spinal Cord Injury: American Spinal
Injury Association. Spinal Cord 1997;35:266-74.
13. Jain AK. Tuberculosis of the spine. Clin Orthop Relat
Res 2007;460:2-3.
14. Garg RK, Raut T, Malhotra HS, et al. Evaluation of
prognostic factors in medically treated patients of
spinal tuberculosis. Rheumatol Int 2013;33:3009-15.
15. Kramer N, Rosenstein ED. Rheumatologic manifes-
tations of tuberculosis. Bull Rheum Dis 1997;46:5-8.
16. Van der Spoel van Dijk A, A MC, Botha PL, Shipley
JA, Kapnoudhis MA, Beukes CA. The diagnosis of
skeletal tuberculosis by polymerase chain reaction.
Cent Afr J Med 2000;46:144-9.
17. Agrawal V, Patgaonkar PR, Nagariya SP. Tuberculosis
of spine. J Craniovertebr Junction Spine 2010;1:74-
85.
18. Garg RK, Somvanshi DS. Spinal tuberculosis: a re-
view. J Spinal Cord Med 2011;34:440-54.
19. Polley P, Dunn R. Noncontiguous spinal tuberculo-
sis: incidence and management. Eur Spine J 2009;18:
1096-101.